State Legislator Reveals AI Regulation Bill Draft

State Rep. Giovanni Capriglione (R-98) has released an initial draft of a comprehensive regulatory framework for artificial intelligence to be introduced in the upcoming 89th legislative session.

The Innovation and Technology Caucus of the Texas Legislature (IT Caucus) chair revealed the regulation bill, cited as the Texas Responsible AI Governance Act, on Monday in an email to stakeholders.

The 41-page bill incorporates recommendations received by the IT Caucus as a result of a request for information released in January, such as limiting regulatory frameworks to high-risk AI systems with the potential to affect individual rights. Low-risk AI systems and models used exclusively for research or testing are exempt under the bill’s “sandbox program exception” to foster innovation in the state.

The framework specifically defines an AI system as one capable of “perceiving an environment through data acquisition and processing and interpreting the derived information to take an action or actions or to imitate intelligent behavior given a specific goal” and “learning and adapting behavior by analyzing how the environment is affected by prior actions.”

Developers, distributors and deployers of high-risk automated decision-making systems (HR ADMS) are subject to regulation under the drafted legislation.

Prior to deployment, developers will be required to provide a statement of intended uses to deployers and submit a report detailing any known system limitations; the type of data used to train the system; the data governance measures taken during training; and any known risks of algorithmic discrimination, unlawful use of personal data or coercion of human behavior resulting from its intended use.

Distributors will be required to recall and/or disable any HR ADMS they have reason to consider as not in compliance with the framework.

For their part, deployers are expected to do the same, immediately suspending the use of any high-risk AI systems until they are brought into compliance.

Additionally, deployers must assign qualified human oversight to the system and complete a semiannual impact assessment detailing the intended use cases of the system, an analysis of known or foreseeable risks of algorithmic discrimination and descriptions of data processed as inputs, produced outputs, data used to customize the system, evaluation metrics, transparency measures taken, user safeguards and cybersecurity measures.

Deployers and developers dealing in systems intended to interact directly with consumers will be required to inform consumers that they are interacting with artificial intelligence and disclose the purpose of the system, that the system can make a consequential decision affecting the consumer, the nature of potential decisions and the factors involved in a decision’s formation.

Under the proposed legislation, the use of AI for the manipulation of human behavior to circumvent free will, exploiting vulnerabilities, social scoring, untargeted facial recognition, emotional recognition, biometric categorization and sexual abuse material would be strictly prohibited.

The bill also includes a subchapter detailing the establishment of an AI workforce development grant program by the Texas Workforce Commission with the intent to upskill Texas workers and expand career programs focused on AI skill development.

In his email, Capriglione requested stakeholder feedback by Nov. 18. Should the bill be passed, it would take effect on Sept. 1, 2025.